Semantic Search

The /v1/search/semantic endpoint (API) takes a text query, typically entered by your end user, and semantic results are returned with their IDs and relevance score. The text query can be short keywords, longer natural language, or anything in between.

📘

Vantage Managed Embeddings (VME) only!

Semantic Search is only available for Vantage Managed Embeddings, where you have an LLM provider configured for your collection.

text

Text that you would like to find the closest semantic matches for in your collection_id. This text is sent to your embedding provider and then a search is performed to find the most relevant results.

👍

Standard Search Options apply

Filtering, pagination, accuracy etc all apply to this endpoint

The example JSON below uses this endpoint to tell the Vantage platform what to search and return. The text field is your search query. It is passed to the LLM that you configured for the specified collection. This creates an embedding that the Vantage platform uses in its semantic similarity search algorithm. The rest of the JSON is covered in the Search Options section below.

{
  "text": "Looking for great summer brunch outfit for Montauk!",
  "request_id": 333666, 
  "collection": {
    "account_id": "docs-account",
    "collection_id": "docs-sample-collection",
    "accuracy": "0.2"
  },
  "filter": {
    "boolean_filter": ""
  },
  "pagination": {
    "page": 0,
    "count": 40
  }
}

📘

Reference Guide for Semantic Search